Skip to content

BACKLOG.md

Latest commit

 

History

History
106 lines (79 loc) · 4.42 KB

BACKLOG.md

File metadata and controls

106 lines (79 loc) · 4.42 KB
Raw

CodeScribe Backlog & Roadmap

Last updated: 2026-02-07

✅ Completed Features

Recording Modes

Feature Status Files
Hold Mode (Fn = Raw) ✅ app/controller/, app/os/hotkeys.rs
Assistive Mode (Fn+Shift = AI) ✅ app/controller/, app/os/hotkeys.rs
Toggle Mode (Double Option) ✅ app/controller/, app/os/hotkeys.rs
VAD Auto-Stop (5s silence) ✅ audio/recorder.rs

Voice Chat UI (Mission Control)

Feature Status Files
Split panel layout (60/40) ✅ app/ui/voice_chat/mod.rs
Chat bubbles (user/assistant) ✅ app/ui/voice_chat/mod.rs
Streaming AI responses ✅ app/ui/voice_chat/api.rs
Transcriptions tab ✅ app/ui/voice_chat/handlers.rs
Settings window ✅ app/ui/settings/
Attachments in chat ✅ app/ui/voice_chat/handlers.rs
Auto-send toggle ✅ app/ui/voice_chat/state.rs
Collapsible right panel ✅ app/ui/voice_chat/mod.rs

Infrastructure

Feature Status Files
Runtime Whisper model lookup ✅ core/stt/whisper/, tests/support/e2e_stt_matrix.rs
Streaming transcription (Whisper Live) ✅ core/audio/streaming_recorder.rs
IPC Server (Unix socket) ✅ app/ipc/server.rs
Quality Loop (self-improvement) ✅ core/quality/quality_loop.rs
Quality Reports (batch analysis) ✅ core/quality/quality_report.rs
CodeScribe Core separation ✅ core/
Tray app with submenus ✅ app/ui/tray/

📋 Planned Features

1. Tauri GUI (Voice Lab)

  • Status: 📋 Not started
  • Goal: Standalone GUI app for voice training and settings
  • Architecture: Tauri + Leptos WASM, imports codescribe-core
  • Features:
    • Voice Lab (record/playback/compare)
    • Teacher mode (side-by-side correction)
    • Visual settings editor
  • Priority: Low (current overlay covers most needs)

2. TTS Integration

  • Status: 📋 Not started
  • Goal: Text-to-Speech for assistive mode responses
  • Integration: Via Libraxis Qube Protocol — <tts> tags in SSE stream
  • Dependency: Requires Libraxis Qube Protocol implementation

3. Libraxis Qube Protocol

  • Status: 📋 Conceptual (docs/future/ARCHITECTURE_VISION.md)
  • Goal: WebSocket-based "Single Stream" architecture
  • Key Concepts:
    • Central orchestrator (localhost or remote Dragon)
    • Tag-based demuxing (<speak>, <artifact>, <ui_message>)
    • Audio streaming over WebSocket
  • Priority: Low (current REST + SSE sufficient)

🔧 Technical Debt

Item Priority Notes
Split legacy voice chat monolith ✅ Done app/ui/voice_chat/*
Split controller.rs (<1000 LOC) ✅ Done 4 modules created
Decouple Settings from overlay ✅ Done Separate settings window
Update lexicon (Roost→Rust, etc.) CRITICAL assets/programming.jsonl

📊 Metrics

Metric Value
Total Rust LOC ~84,500
core/ ~38,000 LOC
app/ ~37,000 LOC
tests/ ~6,800 LOC
Whisper packaging Embedded-first

Related Documentation:

Created by M&K (c)2026 VetCoders